Merge branch 'rr/test-3200-style' into maint
[git/git.git] / dir.c
CommitLineData
453ec4bd
LT
1/*
2 * This handles recursive filename detection with exclude
3 * files, index knowledge etc..
4 *
95a68344
AS
5 * See Documentation/technical/api-directory-listing.txt
6 *
453ec4bd
LT
7 * Copyright (C) Linus Torvalds, 2005-2006
8 * Junio Hamano, 2005-2006
9 */
453ec4bd
LT
10#include "cache.h"
11#include "dir.h"
09595258 12#include "refs.h"
237ec6e4 13#include "wildmatch.h"
453ec4bd 14
9fc42d60
LT
15struct path_simplify {
16 int len;
17 const char *path;
18};
19
dba2e203 20static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
09595258 21 int check_only, const struct path_simplify *simplify);
caa6b782 22static int get_dtype(struct dirent *de, const char *path, int len);
09595258 23
8cf2a84e
JJ
24/* helper string functions with support for the ignore_case flag */
25int strcmp_icase(const char *a, const char *b)
26{
27 return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
28}
29
30int strncmp_icase(const char *a, const char *b, size_t count)
31{
32 return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
33}
34
35int fnmatch_icase(const char *pattern, const char *string, int flags)
36{
37 return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
38}
39
5d74762d
NTND
40inline int git_fnmatch(const char *pattern, const char *string,
41 int flags, int prefix)
42{
43 int fnm_flags = 0;
44 if (flags & GFNM_PATHNAME)
45 fnm_flags |= FNM_PATHNAME;
46 if (prefix > 0) {
47 if (strncmp(pattern, string, prefix))
48 return FNM_NOMATCH;
49 pattern += prefix;
50 string += prefix;
51 }
8c6abbcd
NTND
52 if (flags & GFNM_ONESTAR) {
53 int pattern_len = strlen(++pattern);
54 int string_len = strlen(string);
55 return string_len < pattern_len ||
56 strcmp(pattern,
57 string + string_len - pattern_len);
58 }
5d74762d
NTND
59 return fnmatch(pattern, string, fnm_flags);
60}
61
0b6e56df
JH
62static int fnmatch_icase_mem(const char *pattern, int patternlen,
63 const char *string, int stringlen,
64 int flags)
65{
66 int match_status;
67 struct strbuf pat_buf = STRBUF_INIT;
68 struct strbuf str_buf = STRBUF_INIT;
69 const char *use_pat = pattern;
70 const char *use_str = string;
71
72 if (pattern[patternlen]) {
73 strbuf_add(&pat_buf, pattern, patternlen);
74 use_pat = pat_buf.buf;
75 }
76 if (string[stringlen]) {
77 strbuf_add(&str_buf, string, stringlen);
78 use_str = str_buf.buf;
79 }
80
6466fbbe
JH
81 if (ignore_case)
82 flags |= WM_CASEFOLD;
83 match_status = wildmatch(use_pat, use_str, flags, NULL);
0b6e56df
JH
84
85 strbuf_release(&pat_buf);
86 strbuf_release(&str_buf);
87
88 return match_status;
89}
90
f950eb95 91static size_t common_prefix_len(const char **pathspec)
3c6a370b 92{
4a085b16
JH
93 const char *n, *first;
94 size_t max = 0;
823ab40f 95 int literal = limit_pathspec_to_literal();
3c6a370b
LT
96
97 if (!pathspec)
4a085b16
JH
98 return max;
99
100 first = *pathspec;
101 while ((n = *pathspec++)) {
102 size_t i, len = 0;
103 for (i = 0; first == n || i < max; i++) {
104 char c = n[i];
823ab40f 105 if (!c || c != first[i] || (!literal && is_glob_special(c)))
4a085b16
JH
106 break;
107 if (c == '/')
108 len = i + 1;
109 }
110 if (first == n || len < max) {
111 max = len;
112 if (!max)
113 break;
114 }
3c6a370b 115 }
4a085b16 116 return max;
3c6a370b
LT
117}
118
f950eb95
CB
119/*
120 * Returns a copy of the longest leading path common among all
121 * pathspecs.
122 */
123char *common_prefix(const char **pathspec)
124{
125 unsigned long len = common_prefix_len(pathspec);
126
127 return len ? xmemdupz(*pathspec, len) : NULL;
128}
129
1d8842d9
LT
130int fill_directory(struct dir_struct *dir, const char **pathspec)
131{
4a085b16 132 size_t len;
1d8842d9
LT
133
134 /*
135 * Calculate common prefix for the pathspec, and
136 * use that to optimize the directory walk
137 */
4a085b16 138 len = common_prefix_len(pathspec);
1d8842d9
LT
139
140 /* Read the directory and prune it */
2b189435 141 read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
dba2e203 142 return len;
1d8842d9
LT
143}
144
bc96cc87
NTND
145int within_depth(const char *name, int namelen,
146 int depth, int max_depth)
147{
148 const char *cp = name, *cpe = name + namelen;
149
150 while (cp < cpe) {
151 if (*cp++ != '/')
152 continue;
153 depth++;
154 if (depth > max_depth)
155 return 0;
156 }
157 return 1;
158}
159
e813d50e 160/*
2c5b0115 161 * Does 'match' match the given name?
e813d50e
JH
162 * A match is found if
163 *
164 * (1) the 'match' string is leading directory of 'name', or
165 * (2) the 'match' string is a wildcard and matches 'name', or
166 * (3) the 'match' string is exactly the same as 'name'.
167 *
168 * and the return value tells which case it was.
169 *
170 * It returns 0 when there is no match.
171 */
3c6a370b
LT
172static int match_one(const char *match, const char *name, int namelen)
173{
174 int matchlen;
823ab40f 175 int literal = limit_pathspec_to_literal();
3c6a370b
LT
176
177 /* If the match was just the prefix, we matched */
88ea8112 178 if (!*match)
e813d50e 179 return MATCHED_RECURSIVELY;
3c6a370b 180
21444f18
JJ
181 if (ignore_case) {
182 for (;;) {
183 unsigned char c1 = tolower(*match);
184 unsigned char c2 = tolower(*name);
823ab40f 185 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
186 break;
187 if (c1 != c2)
188 return 0;
189 match++;
190 name++;
191 namelen--;
192 }
193 } else {
194 for (;;) {
195 unsigned char c1 = *match;
196 unsigned char c2 = *name;
823ab40f 197 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
198 break;
199 if (c1 != c2)
200 return 0;
201 match++;
202 name++;
203 namelen--;
204 }
88ea8112
LT
205 }
206
3c6a370b
LT
207 /*
208 * If we don't match the matchstring exactly,
209 * we need to match by fnmatch
210 */
88ea8112 211 matchlen = strlen(match);
823ab40f
JK
212 if (strncmp_icase(match, name, matchlen)) {
213 if (literal)
214 return 0;
21444f18 215 return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
823ab40f 216 }
3c6a370b 217
f2d0df71 218 if (namelen == matchlen)
e813d50e
JH
219 return MATCHED_EXACTLY;
220 if (match[matchlen-1] == '/' || name[matchlen] == '/')
221 return MATCHED_RECURSIVELY;
222 return 0;
3c6a370b
LT
223}
224
e813d50e 225/*
52ed1894
AS
226 * Given a name and a list of pathspecs, returns the nature of the
227 * closest (i.e. most specific) match of the name to any of the
228 * pathspecs.
229 *
230 * The caller typically calls this multiple times with the same
231 * pathspec and seen[] array but with different name/namelen
232 * (e.g. entries from the index) and is interested in seeing if and
233 * how each pathspec matches all the names it calls this function
234 * with. A mark is left in the seen[] array for each pathspec element
235 * indicating the closest type of match that element achieved, so if
236 * seen[n] remains zero after multiple invocations, that means the nth
237 * pathspec did not match any names, which could indicate that the
238 * user mistyped the nth pathspec.
e813d50e 239 */
0b50922a
CB
240int match_pathspec(const char **pathspec, const char *name, int namelen,
241 int prefix, char *seen)
3c6a370b 242{
0b50922a
CB
243 int i, retval = 0;
244
245 if (!pathspec)
246 return 1;
3c6a370b
LT
247
248 name += prefix;
249 namelen -= prefix;
250
0b50922a 251 for (i = 0; pathspec[i] != NULL; i++) {
e813d50e 252 int how;
0b50922a
CB
253 const char *match = pathspec[i] + prefix;
254 if (seen && seen[i] == MATCHED_EXACTLY)
3c6a370b 255 continue;
e813d50e
JH
256 how = match_one(match, name, namelen);
257 if (how) {
258 if (retval < how)
259 retval = how;
0b50922a
CB
260 if (seen && seen[i] < how)
261 seen[i] = how;
3c6a370b
LT
262 }
263 }
264 return retval;
265}
266
61cf2820
NTND
267/*
268 * Does 'match' match the given name?
269 * A match is found if
270 *
271 * (1) the 'match' string is leading directory of 'name', or
272 * (2) the 'match' string is a wildcard and matches 'name', or
273 * (3) the 'match' string is exactly the same as 'name'.
274 *
275 * and the return value tells which case it was.
276 *
277 * It returns 0 when there is no match.
278 */
279static int match_pathspec_item(const struct pathspec_item *item, int prefix,
280 const char *name, int namelen)
281{
282 /* name/namelen has prefix cut off by caller */
283 const char *match = item->match + prefix;
284 int matchlen = item->len - prefix;
285
286 /* If the match was just the prefix, we matched */
287 if (!*match)
288 return MATCHED_RECURSIVELY;
289
290 if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
291 if (matchlen == namelen)
292 return MATCHED_EXACTLY;
293
294 if (match[matchlen-1] == '/' || name[matchlen] == '/')
295 return MATCHED_RECURSIVELY;
296 }
297
5d74762d 298 if (item->nowildcard_len < item->len &&
8c6abbcd
NTND
299 !git_fnmatch(match, name,
300 item->flags & PATHSPEC_ONESTAR ? GFNM_ONESTAR : 0,
301 item->nowildcard_len - prefix))
61cf2820
NTND
302 return MATCHED_FNMATCH;
303
304 return 0;
305}
306
307/*
52ed1894
AS
308 * Given a name and a list of pathspecs, returns the nature of the
309 * closest (i.e. most specific) match of the name to any of the
310 * pathspecs.
311 *
312 * The caller typically calls this multiple times with the same
313 * pathspec and seen[] array but with different name/namelen
314 * (e.g. entries from the index) and is interested in seeing if and
315 * how each pathspec matches all the names it calls this function
316 * with. A mark is left in the seen[] array for each pathspec element
317 * indicating the closest type of match that element achieved, so if
318 * seen[n] remains zero after multiple invocations, that means the nth
319 * pathspec did not match any names, which could indicate that the
320 * user mistyped the nth pathspec.
61cf2820
NTND
321 */
322int match_pathspec_depth(const struct pathspec *ps,
323 const char *name, int namelen,
324 int prefix, char *seen)
325{
326 int i, retval = 0;
327
328 if (!ps->nr) {
329 if (!ps->recursive || ps->max_depth == -1)
330 return MATCHED_RECURSIVELY;
331
332 if (within_depth(name, namelen, 0, ps->max_depth))
333 return MATCHED_EXACTLY;
334 else
335 return 0;
336 }
337
338 name += prefix;
339 namelen -= prefix;
340
341 for (i = ps->nr - 1; i >= 0; i--) {
342 int how;
343 if (seen && seen[i] == MATCHED_EXACTLY)
344 continue;
345 how = match_pathspec_item(ps->items+i, prefix, name, namelen);
346 if (ps->recursive && ps->max_depth != -1 &&
347 how && how != MATCHED_FNMATCH) {
348 int len = ps->items[i].len;
349 if (name[len] == '/')
350 len++;
351 if (within_depth(name+len, namelen-len, 0, ps->max_depth))
352 how = MATCHED_EXACTLY;
353 else
354 how = 0;
355 }
356 if (how) {
357 if (retval < how)
358 retval = how;
359 if (seen && seen[i] < how)
360 seen[i] = how;
361 }
362 }
363 return retval;
364}
365
fcd631ed
NTND
366/*
367 * Return the length of the "simple" part of a path match limiter.
368 */
369static int simple_length(const char *match)
370{
371 int len = -1;
372
373 for (;;) {
374 unsigned char c = *match++;
375 len++;
376 if (c == '\0' || is_glob_special(c))
377 return len;
378 }
379}
380
68492fc7
LK
381static int no_wildcard(const char *string)
382{
fcd631ed 383 return string[simple_length(string)] == '\0';
68492fc7
LK
384}
385
82dce998
NTND
386void parse_exclude_pattern(const char **pattern,
387 int *patternlen,
388 int *flags,
389 int *nowildcardlen)
84460eec
NTND
390{
391 const char *p = *pattern;
392 size_t i, len;
393
394 *flags = 0;
395 if (*p == '!') {
396 *flags |= EXC_FLAG_NEGATIVE;
397 p++;
398 }
399 len = strlen(p);
400 if (len && p[len - 1] == '/') {
401 len--;
402 *flags |= EXC_FLAG_MUSTBEDIR;
403 }
404 for (i = 0; i < len; i++) {
405 if (p[i] == '/')
406 break;
407 }
408 if (i == len)
409 *flags |= EXC_FLAG_NODIR;
410 *nowildcardlen = simple_length(p);
411 /*
412 * we should have excluded the trailing slash from 'p' too,
413 * but that's one more allocation. Instead just make sure
414 * nowildcardlen does not exceed real patternlen
415 */
416 if (*nowildcardlen > len)
417 *nowildcardlen = len;
418 if (*p == '*' && no_wildcard(p + 1))
419 *flags |= EXC_FLAG_ENDSWITH;
420 *pattern = p;
421 *patternlen = len;
422}
423
453ec4bd 424void add_exclude(const char *string, const char *base,
c04318e4 425 int baselen, struct exclude_list *el, int srcpos)
453ec4bd 426{
d6b8fc30 427 struct exclude *x;
84460eec
NTND
428 int patternlen;
429 int flags;
430 int nowildcardlen;
453ec4bd 431
84460eec
NTND
432 parse_exclude_pattern(&string, &patternlen, &flags, &nowildcardlen);
433 if (flags & EXC_FLAG_MUSTBEDIR) {
d6b8fc30 434 char *s;
84460eec 435 x = xmalloc(sizeof(*x) + patternlen + 1);
4b25d091 436 s = (char *)(x+1);
84460eec
NTND
437 memcpy(s, string, patternlen);
438 s[patternlen] = '\0';
d6b8fc30 439 x->pattern = s;
d6b8fc30
JH
440 } else {
441 x = xmalloc(sizeof(*x));
442 x->pattern = string;
443 }
84460eec
NTND
444 x->patternlen = patternlen;
445 x->nowildcardlen = nowildcardlen;
453ec4bd
LT
446 x->base = base;
447 x->baselen = baselen;
d6b8fc30 448 x->flags = flags;
c04318e4 449 x->srcpos = srcpos;
840fc334
AS
450 ALLOC_GROW(el->excludes, el->nr + 1, el->alloc);
451 el->excludes[el->nr++] = x;
c04318e4 452 x->el = el;
453ec4bd
LT
453}
454
c28b3d6e
NTND
455static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
456{
457 int pos, len;
458 unsigned long sz;
459 enum object_type type;
460 void *data;
461 struct index_state *istate = &the_index;
462
463 len = strlen(path);
464 pos = index_name_pos(istate, path, len);
465 if (pos < 0)
466 return NULL;
467 if (!ce_skip_worktree(istate->cache[pos]))
468 return NULL;
469 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
470 if (!data || type != OBJ_BLOB) {
471 free(data);
472 return NULL;
473 }
474 *size = xsize_t(sz);
475 return data;
476}
477
f6198812
AS
478/*
479 * Frees memory within el which was allocated for exclude patterns and
480 * the file buffer. Does not free el itself.
481 */
482void clear_exclude_list(struct exclude_list *el)
0fd0e241
NTND
483{
484 int i;
485
486 for (i = 0; i < el->nr; i++)
487 free(el->excludes[i]);
488 free(el->excludes);
c082df24 489 free(el->filebuf);
0fd0e241
NTND
490
491 el->nr = 0;
492 el->excludes = NULL;
c082df24 493 el->filebuf = NULL;
0fd0e241
NTND
494}
495
cb097534
NTND
496int add_excludes_from_file_to_list(const char *fname,
497 const char *base,
498 int baselen,
840fc334 499 struct exclude_list *el,
cb097534 500 int check_index)
453ec4bd 501{
c470701a 502 struct stat st;
c04318e4 503 int fd, i, lineno = 1;
9d14017a 504 size_t size = 0;
453ec4bd
LT
505 char *buf, *entry;
506
507 fd = open(fname, O_RDONLY);
c28b3d6e 508 if (fd < 0 || fstat(fd, &st) < 0) {
69660731 509 if (errno != ENOENT)
55b38a48 510 warn_on_inaccessible(fname);
c28b3d6e
NTND
511 if (0 <= fd)
512 close(fd);
513 if (!check_index ||
514 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
515 return -1;
45d76f17
NTND
516 if (size == 0) {
517 free(buf);
518 return 0;
519 }
520 if (buf[size-1] != '\n') {
521 buf = xrealloc(buf, size+1);
522 buf[size++] = '\n';
523 }
453ec4bd 524 }
c28b3d6e
NTND
525 else {
526 size = xsize_t(st.st_size);
527 if (size == 0) {
528 close(fd);
529 return 0;
530 }
45d76f17 531 buf = xmalloc(size+1);
c28b3d6e 532 if (read_in_full(fd, buf, size) != size) {
45d76f17 533 free(buf);
c28b3d6e
NTND
534 close(fd);
535 return -1;
536 }
45d76f17 537 buf[size++] = '\n';
c28b3d6e 538 close(fd);
6ba78238 539 }
453ec4bd 540
c082df24 541 el->filebuf = buf;
453ec4bd 542 entry = buf;
45d76f17
NTND
543 for (i = 0; i < size; i++) {
544 if (buf[i] == '\n') {
453ec4bd
LT
545 if (entry != buf + i && entry[0] != '#') {
546 buf[i - (i && buf[i-1] == '\r')] = 0;
c04318e4 547 add_exclude(entry, base, baselen, el, lineno);
453ec4bd 548 }
c04318e4 549 lineno++;
453ec4bd
LT
550 entry = buf + i + 1;
551 }
552 }
553 return 0;
453ec4bd
LT
554}
555
c04318e4
AS
556struct exclude_list *add_exclude_list(struct dir_struct *dir,
557 int group_type, const char *src)
c082df24
AS
558{
559 struct exclude_list *el;
560 struct exclude_list_group *group;
561
562 group = &dir->exclude_list_group[group_type];
563 ALLOC_GROW(group->el, group->nr + 1, group->alloc);
564 el = &group->el[group->nr++];
565 memset(el, 0, sizeof(*el));
c04318e4 566 el->src = src;
c082df24
AS
567 return el;
568}
569
570/*
571 * Used to set up core.excludesfile and .git/info/exclude lists.
572 */
453ec4bd
LT
573void add_excludes_from_file(struct dir_struct *dir, const char *fname)
574{
c082df24 575 struct exclude_list *el;
c04318e4 576 el = add_exclude_list(dir, EXC_FILE, fname);
c082df24 577 if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
453ec4bd
LT
578 die("cannot use %s as an exclude file", fname);
579}
580
95a68344
AS
581/*
582 * Loads the per-directory exclude list for the substring of base
583 * which has a char length of baselen.
584 */
63d285c8 585static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
453ec4bd 586{
c082df24 587 struct exclude_list_group *group;
63d285c8
JH
588 struct exclude_list *el;
589 struct exclude_stack *stk = NULL;
590 int current;
591
592 if ((!dir->exclude_per_dir) ||
593 (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
594 return; /* too long a path -- ignore */
595
c082df24
AS
596 group = &dir->exclude_list_group[EXC_DIRS];
597
598 /* Pop the exclude lists from the EXCL_DIRS exclude_list_group
599 * which originate from directories not in the prefix of the
600 * path being checked. */
63d285c8
JH
601 while ((stk = dir->exclude_stack) != NULL) {
602 if (stk->baselen <= baselen &&
603 !strncmp(dir->basebuf, base, stk->baselen))
604 break;
c082df24 605 el = &group->el[dir->exclude_stack->exclude_ix];
63d285c8 606 dir->exclude_stack = stk->prev;
c04318e4 607 free((char *)el->src); /* see strdup() below */
c082df24 608 clear_exclude_list(el);
63d285c8 609 free(stk);
c082df24 610 group->nr--;
453ec4bd 611 }
453ec4bd 612
63d285c8
JH
613 /* Read from the parent directories and push them down. */
614 current = stk ? stk->baselen : -1;
615 while (current < baselen) {
616 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
617 const char *cp;
453ec4bd 618
63d285c8
JH
619 if (current < 0) {
620 cp = base;
621 current = 0;
622 }
623 else {
624 cp = strchr(base + current + 1, '/');
625 if (!cp)
626 die("oops in prep_exclude");
627 cp++;
628 }
629 stk->prev = dir->exclude_stack;
630 stk->baselen = cp - base;
63d285c8
JH
631 memcpy(dir->basebuf + current, base + current,
632 stk->baselen - current);
633 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
c04318e4
AS
634 /*
635 * dir->basebuf gets reused by the traversal, but we
636 * need fname to remain unchanged to ensure the src
637 * member of each struct exclude correctly
638 * back-references its source file. Other invocations
639 * of add_exclude_list provide stable strings, so we
640 * strdup() and free() here in the caller.
641 */
642 el = add_exclude_list(dir, EXC_DIRS, strdup(dir->basebuf));
c082df24 643 stk->exclude_ix = group->nr - 1;
cb097534
NTND
644 add_excludes_from_file_to_list(dir->basebuf,
645 dir->basebuf, stk->baselen,
c082df24 646 el, 1);
63d285c8
JH
647 dir->exclude_stack = stk;
648 current = stk->baselen;
649 }
650 dir->basebuf[baselen] = '\0';
453ec4bd
LT
651}
652
82dce998
NTND
653int match_basename(const char *basename, int basenamelen,
654 const char *pattern, int prefix, int patternlen,
655 int flags)
593cb880
NTND
656{
657 if (prefix == patternlen) {
0b6e56df
JH
658 if (patternlen == basenamelen &&
659 !strncmp_icase(pattern, basename, basenamelen))
593cb880
NTND
660 return 1;
661 } else if (flags & EXC_FLAG_ENDSWITH) {
0b6e56df 662 /* "*literal" matching against "fooliteral" */
593cb880 663 if (patternlen - 1 <= basenamelen &&
0b6e56df
JH
664 !strncmp_icase(pattern + 1,
665 basename + basenamelen - (patternlen - 1),
666 patternlen - 1))
593cb880
NTND
667 return 1;
668 } else {
0b6e56df
JH
669 if (fnmatch_icase_mem(pattern, patternlen,
670 basename, basenamelen,
671 0) == 0)
593cb880
NTND
672 return 1;
673 }
674 return 0;
675}
676
82dce998
NTND
677int match_pathname(const char *pathname, int pathlen,
678 const char *base, int baselen,
679 const char *pattern, int prefix, int patternlen,
680 int flags)
b5592632
NTND
681{
682 const char *name;
683 int namelen;
684
685 /*
686 * match with FNM_PATHNAME; the pattern has base implicitly
687 * in front of it.
688 */
689 if (*pattern == '/') {
690 pattern++;
982ac873 691 patternlen--;
b5592632
NTND
692 prefix--;
693 }
694
695 /*
696 * baselen does not count the trailing slash. base[] may or
697 * may not end with a trailing slash though.
698 */
699 if (pathlen < baselen + 1 ||
700 (baselen && pathname[baselen] != '/') ||
701 strncmp_icase(pathname, base, baselen))
702 return 0;
703
704 namelen = baselen ? pathlen - baselen - 1 : pathlen;
705 name = pathname + pathlen - namelen;
706
707 if (prefix) {
708 /*
709 * if the non-wildcard part is longer than the
710 * remaining pathname, surely it cannot match.
711 */
712 if (prefix > namelen)
713 return 0;
714
715 if (strncmp_icase(pattern, name, prefix))
716 return 0;
717 pattern += prefix;
ab3aebc1 718 patternlen -= prefix;
b5592632
NTND
719 name += prefix;
720 namelen -= prefix;
ab3aebc1
JK
721
722 /*
723 * If the whole pattern did not have a wildcard,
724 * then our prefix match is all we need; we
725 * do not need to call fnmatch at all.
726 */
727 if (!patternlen && !namelen)
728 return 1;
b5592632
NTND
729 }
730
ab3aebc1
JK
731 return fnmatch_icase_mem(pattern, patternlen,
732 name, namelen,
6466fbbe 733 WM_PATHNAME) == 0;
b5592632
NTND
734}
735
578cd7c3
AS
736/*
737 * Scan the given exclude list in reverse to see whether pathname
738 * should be ignored. The first match (i.e. the last on the list), if
739 * any, determines the fate. Returns the exclude_list element which
740 * matched, or NULL for undecided.
453ec4bd 741 */
578cd7c3
AS
742static struct exclude *last_exclude_matching_from_list(const char *pathname,
743 int pathlen,
744 const char *basename,
745 int *dtype,
746 struct exclude_list *el)
453ec4bd
LT
747{
748 int i;
749
35a94d44 750 if (!el->nr)
578cd7c3 751 return NULL; /* undefined */
d6b8fc30 752
35a94d44
NTND
753 for (i = el->nr - 1; 0 <= i; i--) {
754 struct exclude *x = el->excludes[i];
b5592632 755 const char *exclude = x->pattern;
b5592632 756 int prefix = x->nowildcardlen;
35a94d44
NTND
757
758 if (x->flags & EXC_FLAG_MUSTBEDIR) {
759 if (*dtype == DT_UNKNOWN)
760 *dtype = get_dtype(NULL, pathname, pathlen);
761 if (*dtype != DT_DIR)
762 continue;
763 }
d6b8fc30 764
35a94d44 765 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
766 if (match_basename(basename,
767 pathlen - (basename - pathname),
768 exclude, prefix, x->patternlen,
769 x->flags))
578cd7c3 770 return x;
35a94d44
NTND
771 continue;
772 }
773
b5592632
NTND
774 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
775 if (match_pathname(pathname, pathlen,
776 x->base, x->baselen ? x->baselen - 1 : 0,
777 exclude, prefix, x->patternlen, x->flags))
578cd7c3 778 return x;
453ec4bd 779 }
578cd7c3
AS
780 return NULL; /* undecided */
781}
782
783/*
784 * Scan the list and let the last match determine the fate.
785 * Return 1 for exclude, 0 for include and -1 for undecided.
786 */
787int is_excluded_from_list(const char *pathname,
788 int pathlen, const char *basename, int *dtype,
789 struct exclude_list *el)
790{
791 struct exclude *exclude;
792 exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
793 if (exclude)
794 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
795 return -1; /* undecided */
796}
797
f4cd69a6
AS
798/*
799 * Loads the exclude lists for the directory containing pathname, then
800 * scans all exclude lists to determine whether pathname is excluded.
801 * Returns the exclude_list element which matched, or NULL for
802 * undecided.
803 */
804static struct exclude *last_exclude_matching(struct dir_struct *dir,
805 const char *pathname,
806 int *dtype_p)
453ec4bd
LT
807{
808 int pathlen = strlen(pathname);
c082df24
AS
809 int i, j;
810 struct exclude_list_group *group;
f4cd69a6 811 struct exclude *exclude;
68492fc7
LK
812 const char *basename = strrchr(pathname, '/');
813 basename = (basename) ? basename+1 : pathname;
453ec4bd 814
63d285c8 815 prep_exclude(dir, pathname, basename-pathname);
c082df24
AS
816
817 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
818 group = &dir->exclude_list_group[i];
819 for (j = group->nr - 1; j >= 0; j--) {
820 exclude = last_exclude_matching_from_list(
821 pathname, pathlen, basename, dtype_p,
822 &group->el[j]);
823 if (exclude)
824 return exclude;
825 }
453ec4bd 826 }
f4cd69a6
AS
827 return NULL;
828}
829
830/*
831 * Loads the exclude lists for the directory containing pathname, then
832 * scans all exclude lists to determine whether pathname is excluded.
833 * Returns 1 if true, otherwise 0.
834 */
835static int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
836{
837 struct exclude *exclude =
838 last_exclude_matching(dir, pathname, dtype_p);
839 if (exclude)
840 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
841 return 0;
842}
843
eb41775e
JH
844void path_exclude_check_init(struct path_exclude_check *check,
845 struct dir_struct *dir)
846{
847 check->dir = dir;
a35341a8 848 check->exclude = NULL;
eb41775e
JH
849 strbuf_init(&check->path, 256);
850}
851
852void path_exclude_check_clear(struct path_exclude_check *check)
853{
854 strbuf_release(&check->path);
855}
856
93921b07 857/*
a35341a8
AS
858 * For each subdirectory in name, starting with the top-most, checks
859 * to see if that subdirectory is excluded, and if so, returns the
860 * corresponding exclude structure. Otherwise, checks whether name
861 * itself (which is presumably a file) is excluded.
93921b07
JH
862 *
863 * A path to a directory known to be excluded is left in check->path to
864 * optimize for repeated checks for files in the same excluded directory.
865 */
a35341a8
AS
866struct exclude *last_exclude_matching_path(struct path_exclude_check *check,
867 const char *name, int namelen,
868 int *dtype)
eb41775e 869{
782cd4c0 870 int i;
eb41775e 871 struct strbuf *path = &check->path;
a35341a8 872 struct exclude *exclude;
eb41775e 873
782cd4c0
JH
874 /*
875 * we allow the caller to pass namelen as an optimization; it
876 * must match the length of the name, as we eventually call
6d24e7a8 877 * is_excluded() on the whole name string.
782cd4c0
JH
878 */
879 if (namelen < 0)
880 namelen = strlen(name);
881
a35341a8
AS
882 /*
883 * If path is non-empty, and name is equal to path or a
884 * subdirectory of path, name should be excluded, because
885 * it's inside a directory which is already known to be
886 * excluded and was previously left in check->path.
887 */
93921b07 888 if (path->len &&
782cd4c0
JH
889 path->len <= namelen &&
890 !memcmp(name, path->buf, path->len) &&
891 (!name[path->len] || name[path->len] == '/'))
a35341a8 892 return check->exclude;
93921b07 893
eb41775e 894 strbuf_setlen(path, 0);
782cd4c0
JH
895 for (i = 0; name[i]; i++) {
896 int ch = name[i];
eb41775e
JH
897
898 if (ch == '/') {
782cd4c0 899 int dt = DT_DIR;
a35341a8
AS
900 exclude = last_exclude_matching(check->dir,
901 path->buf, &dt);
902 if (exclude) {
903 check->exclude = exclude;
904 return exclude;
905 }
eb41775e
JH
906 }
907 strbuf_addch(path, ch);
908 }
93921b07
JH
909
910 /* An entry in the index; cannot be a directory with subentries */
911 strbuf_setlen(path, 0);
912
a35341a8
AS
913 return last_exclude_matching(check->dir, name, dtype);
914}
915
916/*
917 * Is this name excluded? This is for a caller like show_files() that
918 * do not honor directory hierarchy and iterate through paths that are
919 * possibly in an ignored directory.
920 */
921int is_path_excluded(struct path_exclude_check *check,
922 const char *name, int namelen, int *dtype)
923{
924 struct exclude *exclude =
925 last_exclude_matching_path(check, name, namelen, dtype);
926 if (exclude)
927 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
928 return 0;
eb41775e
JH
929}
930
f3fa1838
JH
931static struct dir_entry *dir_entry_new(const char *pathname, int len)
932{
453ec4bd
LT
933 struct dir_entry *ent;
934
453ec4bd
LT
935 ent = xmalloc(sizeof(*ent) + len + 1);
936 ent->len = len;
937 memcpy(ent->name, pathname, len);
938 ent->name[len] = 0;
4d06f8ac 939 return ent;
453ec4bd
LT
940}
941
159b3212 942static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 943{
a45fb697
AP
944 if (!(dir->flags & DIR_SHOW_IGNORED) &&
945 cache_name_exists(pathname, len, ignore_case))
6815e569
JK
946 return NULL;
947
25fd2f7a 948 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
949 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
950}
951
108da0db 952struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 953{
6e4f981f 954 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
955 return NULL;
956
25fd2f7a 957 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
958 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
959}
960
09595258
LT
961enum exist_status {
962 index_nonexistent = 0,
963 index_directory,
4b05548f 964 index_gitdir
09595258
LT
965};
966
5102c617
JJ
967/*
968 * Do not use the alphabetically stored index to look up
969 * the directory name; instead, use the case insensitive
970 * name hash.
971 */
972static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
973{
974 struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
975 unsigned char endchar;
976
977 if (!ce)
978 return index_nonexistent;
979 endchar = ce->name[len];
980
981 /*
982 * The cache_entry structure returned will contain this dirname
983 * and possibly additional path components.
984 */
985 if (endchar == '/')
986 return index_directory;
987
988 /*
989 * If there are no additional path components, then this cache_entry
990 * represents a submodule. Submodules, despite being directories,
991 * are stored in the cache without a closing slash.
992 */
993 if (!endchar && S_ISGITLINK(ce->ce_mode))
994 return index_gitdir;
995
996 /* This should never be hit, but it exists just in case. */
997 return index_nonexistent;
998}
999
09595258
LT
1000/*
1001 * The index sorts alphabetically by entry name, which
1002 * means that a gitlink sorts as '\0' at the end, while
1003 * a directory (which is defined not as an entry, but as
1004 * the files it contains) will sort with the '/' at the
1005 * end.
1006 */
1007static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 1008{
5102c617
JJ
1009 int pos;
1010
1011 if (ignore_case)
1012 return directory_exists_in_index_icase(dirname, len);
1013
1014 pos = cache_name_pos(dirname, len);
09595258
LT
1015 if (pos < 0)
1016 pos = -pos-1;
1017 while (pos < active_nr) {
1018 struct cache_entry *ce = active_cache[pos++];
1019 unsigned char endchar;
1020
1021 if (strncmp(ce->name, dirname, len))
1022 break;
1023 endchar = ce->name[len];
1024 if (endchar > '/')
1025 break;
1026 if (endchar == '/')
1027 return index_directory;
7a51ed66 1028 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
1029 return index_gitdir;
1030 }
1031 return index_nonexistent;
1032}
1033
1034/*
1035 * When we find a directory when traversing the filesystem, we
1036 * have three distinct cases:
1037 *
1038 * - ignore it
1039 * - see it as a directory
1040 * - recurse into it
1041 *
1042 * and which one we choose depends on a combination of existing
1043 * git index contents and the flags passed into the directory
1044 * traversal routine.
1045 *
1046 * Case 1: If we *already* have entries in the index under that
721ac4ed
AP
1047 * directory name, we recurse into the directory to see all the files,
1048 * unless the directory is excluded and we want to show ignored
1049 * directories
09595258
LT
1050 *
1051 * Case 2: If we *already* have that directory name as a gitlink,
1052 * we always continue to see it as a gitlink, regardless of whether
1053 * there is an actual git directory there or not (it might not
1054 * be checked out as a subproject!)
1055 *
1056 * Case 3: if we didn't have it in the index previously, we
1057 * have a few sub-cases:
1058 *
1059 * (a) if "show_other_directories" is true, we show it as
1060 * just a directory, unless "hide_empty_directories" is
1061 * also true and the directory is empty, in which case
1062 * we just ignore it entirely.
721ac4ed
AP
1063 * if we are looking for ignored directories, look if it
1064 * contains only ignored files to decide if it must be shown as
1065 * ignored or not.
09595258 1066 * (b) if it looks like a git directory, and we don't have
302b9282 1067 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
1068 * as a directory.
1069 * (c) otherwise, we recurse into it.
1070 */
1071enum directory_treatment {
1072 show_directory,
1073 ignore_directory,
4b05548f 1074 recurse_into_directory
09595258
LT
1075};
1076
1077static enum directory_treatment treat_directory(struct dir_struct *dir,
721ac4ed 1078 const char *dirname, int len, int exclude,
09595258
LT
1079 const struct path_simplify *simplify)
1080{
1081 /* The "len-1" is to strip the final '/' */
1082 switch (directory_exists_in_index(dirname, len-1)) {
1083 case index_directory:
721ac4ed
AP
1084 if ((dir->flags & DIR_SHOW_OTHER_DIRECTORIES) && exclude)
1085 break;
1086
09595258
LT
1087 return recurse_into_directory;
1088
1089 case index_gitdir:
7c4c97c0 1090 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
ab22aed3 1091 return ignore_directory;
09595258
LT
1092 return show_directory;
1093
1094 case index_nonexistent:
7c4c97c0 1095 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 1096 break;
7c4c97c0 1097 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
1098 unsigned char sha1[20];
1099 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
1100 return show_directory;
1101 }
1102 return recurse_into_directory;
1103 }
1104
1105 /* This is the "show_other_directories" case */
721ac4ed
AP
1106
1107 /*
1108 * We are looking for ignored files and our directory is not ignored,
1109 * check if it contains only ignored files
1110 */
1111 if ((dir->flags & DIR_SHOW_IGNORED) && !exclude) {
1112 int ignored;
1113 dir->flags &= ~DIR_SHOW_IGNORED;
1114 dir->flags |= DIR_HIDE_EMPTY_DIRECTORIES;
1115 ignored = read_directory_recursive(dir, dirname, len, 1, simplify);
1116 dir->flags &= ~DIR_HIDE_EMPTY_DIRECTORIES;
1117 dir->flags |= DIR_SHOW_IGNORED;
1118
1119 return ignored ? ignore_directory : show_directory;
1120 }
1121 if (!(dir->flags & DIR_SHOW_IGNORED) &&
1122 !(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
09595258 1123 return show_directory;
dba2e203 1124 if (!read_directory_recursive(dir, dirname, len, 1, simplify))
09595258
LT
1125 return ignore_directory;
1126 return show_directory;
453ec4bd
LT
1127}
1128
721ac4ed
AP
1129/*
1130 * Decide what to do when we find a file while traversing the
1131 * filesystem. Mostly two cases:
1132 *
1133 * 1. We are looking for ignored files
1134 * (a) File is ignored, include it
1135 * (b) File is in ignored path, include it
1136 * (c) File is not ignored, exclude it
1137 *
1138 * 2. Other scenarios, include the file if not excluded
1139 *
1140 * Return 1 for exclude, 0 for include.
1141 */
1142static int treat_file(struct dir_struct *dir, struct strbuf *path, int exclude, int *dtype)
1143{
1144 struct path_exclude_check check;
1145 int exclude_file = 0;
1146
1147 if (exclude)
1148 exclude_file = !(dir->flags & DIR_SHOW_IGNORED);
1149 else if (dir->flags & DIR_SHOW_IGNORED) {
a45fb697 1150 /* Always exclude indexed files */
721ac4ed
AP
1151 struct cache_entry *ce = index_name_exists(&the_index,
1152 path->buf, path->len, ignore_case);
1153
1154 if (ce)
1155 return 1;
1156
1157 path_exclude_check_init(&check, dir);
1158
0a9a787f 1159 if (!is_path_excluded(&check, path->buf, path->len, dtype))
721ac4ed
AP
1160 exclude_file = 1;
1161
1162 path_exclude_check_clear(&check);
1163 }
1164
1165 return exclude_file;
1166}
1167
9fc42d60
LT
1168/*
1169 * This is an inexact early pruning of any recursive directory
1170 * reading - if the path cannot possibly be in the pathspec,
1171 * return true, and we'll skip it early.
1172 */
1173static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1174{
1175 if (simplify) {
1176 for (;;) {
1177 const char *match = simplify->path;
1178 int len = simplify->len;
1179
1180 if (!match)
1181 break;
1182 if (len > pathlen)
1183 len = pathlen;
1184 if (!memcmp(path, match, len))
1185 return 0;
1186 simplify++;
1187 }
1188 return 1;
1189 }
1190 return 0;
1191}
1192
29209cbe
JK
1193/*
1194 * This function tells us whether an excluded path matches a
1195 * list of "interesting" pathspecs. That is, whether a path matched
1196 * by any of the pathspecs could possibly be ignored by excluding
1197 * the specified path. This can happen if:
1198 *
1199 * 1. the path is mentioned explicitly in the pathspec
1200 *
1201 * 2. the path is a directory prefix of some element in the
1202 * pathspec
1203 */
1204static int exclude_matches_pathspec(const char *path, int len,
1205 const struct path_simplify *simplify)
e96980ef
JK
1206{
1207 if (simplify) {
1208 for (; simplify->path; simplify++) {
1209 if (len == simplify->len
1210 && !memcmp(path, simplify->path, len))
1211 return 1;
29209cbe
JK
1212 if (len < simplify->len
1213 && simplify->path[len] == '/'
1214 && !memcmp(path, simplify->path, len))
1215 return 1;
e96980ef
JK
1216 }
1217 }
1218 return 0;
1219}
1220
443e061a
LT
1221static int get_index_dtype(const char *path, int len)
1222{
1223 int pos;
1224 struct cache_entry *ce;
1225
1226 ce = cache_name_exists(path, len, 0);
1227 if (ce) {
1228 if (!ce_uptodate(ce))
1229 return DT_UNKNOWN;
1230 if (S_ISGITLINK(ce->ce_mode))
1231 return DT_DIR;
1232 /*
1233 * Nobody actually cares about the
1234 * difference between DT_LNK and DT_REG
1235 */
1236 return DT_REG;
1237 }
1238
1239 /* Try to look it up as a directory */
1240 pos = cache_name_pos(path, len);
1241 if (pos >= 0)
1242 return DT_UNKNOWN;
1243 pos = -pos-1;
1244 while (pos < active_nr) {
1245 ce = active_cache[pos++];
1246 if (strncmp(ce->name, path, len))
1247 break;
1248 if (ce->name[len] > '/')
1249 break;
1250 if (ce->name[len] < '/')
1251 continue;
1252 if (!ce_uptodate(ce))
1253 break; /* continue? */
1254 return DT_DIR;
1255 }
1256 return DT_UNKNOWN;
1257}
1258
caa6b782 1259static int get_dtype(struct dirent *de, const char *path, int len)
07134421 1260{
6831a88a 1261 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
1262 struct stat st;
1263
1264 if (dtype != DT_UNKNOWN)
1265 return dtype;
443e061a
LT
1266 dtype = get_index_dtype(path, len);
1267 if (dtype != DT_UNKNOWN)
1268 return dtype;
1269 if (lstat(path, &st))
07134421
LT
1270 return dtype;
1271 if (S_ISREG(st.st_mode))
1272 return DT_REG;
1273 if (S_ISDIR(st.st_mode))
1274 return DT_DIR;
1275 if (S_ISLNK(st.st_mode))
1276 return DT_LNK;
1277 return dtype;
1278}
1279
53cc5356
JH
1280enum path_treatment {
1281 path_ignored,
1282 path_handled,
4b05548f 1283 path_recurse
53cc5356
JH
1284};
1285
16e2cfa9 1286static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 1287 struct strbuf *path,
16e2cfa9
JH
1288 const struct path_simplify *simplify,
1289 int dtype, struct dirent *de)
53cc5356 1290{
6d24e7a8 1291 int exclude = is_excluded(dir, path->buf, &dtype);
53cc5356 1292 if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
49dc2cc2
RS
1293 && exclude_matches_pathspec(path->buf, path->len, simplify))
1294 dir_add_ignored(dir, path->buf, path->len);
53cc5356
JH
1295
1296 /*
1297 * Excluded? If we don't explicitly want to show
1298 * ignored files, ignore it
1299 */
1300 if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
1301 return path_ignored;
1302
1303 if (dtype == DT_UNKNOWN)
49dc2cc2 1304 dtype = get_dtype(de, path->buf, path->len);
53cc5356 1305
53cc5356
JH
1306 switch (dtype) {
1307 default:
1308 return path_ignored;
1309 case DT_DIR:
49dc2cc2 1310 strbuf_addch(path, '/');
721ac4ed
AP
1311
1312 switch (treat_directory(dir, path->buf, path->len, exclude, simplify)) {
53cc5356 1313 case show_directory:
53cc5356
JH
1314 break;
1315 case recurse_into_directory:
1316 return path_recurse;
1317 case ignore_directory:
1318 return path_ignored;
1319 }
1320 break;
1321 case DT_REG:
1322 case DT_LNK:
721ac4ed
AP
1323 switch (treat_file(dir, path, exclude, &dtype)) {
1324 case 1:
1325 return path_ignored;
1326 default:
1327 break;
1328 }
53cc5356
JH
1329 }
1330 return path_handled;
1331}
1332
16e2cfa9
JH
1333static enum path_treatment treat_path(struct dir_struct *dir,
1334 struct dirent *de,
49dc2cc2 1335 struct strbuf *path,
16e2cfa9 1336 int baselen,
49dc2cc2 1337 const struct path_simplify *simplify)
16e2cfa9
JH
1338{
1339 int dtype;
1340
1341 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
1342 return path_ignored;
49dc2cc2
RS
1343 strbuf_setlen(path, baselen);
1344 strbuf_addstr(path, de->d_name);
1345 if (simplify_away(path->buf, path->len, simplify))
16e2cfa9
JH
1346 return path_ignored;
1347
1348 dtype = DTYPE(de);
49dc2cc2 1349 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1350}
1351
453ec4bd
LT
1352/*
1353 * Read a directory tree. We currently ignore anything but
1354 * directories, regular files and symlinks. That's because git
1355 * doesn't handle them at all yet. Maybe that will change some
1356 * day.
1357 *
1358 * Also, we ignore the name ".git" (even if it is not a directory).
1359 * That likely will not change.
1360 */
53cc5356
JH
1361static int read_directory_recursive(struct dir_struct *dir,
1362 const char *base, int baselen,
1363 int check_only,
1364 const struct path_simplify *simplify)
453ec4bd 1365{
1528d247 1366 DIR *fdir;
453ec4bd 1367 int contents = 0;
02cb6753 1368 struct dirent *de;
bef36921 1369 struct strbuf path = STRBUF_INIT;
453ec4bd 1370
bef36921 1371 strbuf_add(&path, base, baselen);
02cb6753 1372
1528d247
RS
1373 fdir = opendir(path.len ? path.buf : ".");
1374 if (!fdir)
1375 goto out;
1376
02cb6753 1377 while ((de = readdir(fdir)) != NULL) {
bef36921 1378 switch (treat_path(dir, de, &path, baselen, simplify)) {
02cb6753 1379 case path_recurse:
bef36921
JH
1380 contents += read_directory_recursive(dir, path.buf,
1381 path.len, 0,
1382 simplify);
02cb6753
NTND
1383 continue;
1384 case path_ignored:
1385 continue;
1386 case path_handled:
1387 break;
453ec4bd 1388 }
02cb6753
NTND
1389 contents++;
1390 if (check_only)
1528d247
RS
1391 break;
1392 dir_add_name(dir, path.buf, path.len);
453ec4bd 1393 }
02cb6753 1394 closedir(fdir);
1528d247 1395 out:
bef36921 1396 strbuf_release(&path);
453ec4bd
LT
1397
1398 return contents;
1399}
1400
1401static int cmp_name(const void *p1, const void *p2)
1402{
1403 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1404 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1405
1406 return cache_name_compare(e1->name, e1->len,
1407 e2->name, e2->len);
1408}
1409
9fc42d60
LT
1410static struct path_simplify *create_simplify(const char **pathspec)
1411{
1412 int nr, alloc = 0;
1413 struct path_simplify *simplify = NULL;
1414
1415 if (!pathspec)
1416 return NULL;
1417
1418 for (nr = 0 ; ; nr++) {
1419 const char *match;
1420 if (nr >= alloc) {
1421 alloc = alloc_nr(alloc);
1422 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1423 }
1424 match = *pathspec++;
1425 if (!match)
1426 break;
1427 simplify[nr].path = match;
1428 simplify[nr].len = simple_length(match);
1429 }
1430 simplify[nr].path = NULL;
1431 simplify[nr].len = 0;
1432 return simplify;
1433}
1434
1435static void free_simplify(struct path_simplify *simplify)
1436{
8e0f7003 1437 free(simplify);
9fc42d60
LT
1438}
1439
48ffef96
JH
1440static int treat_leading_path(struct dir_struct *dir,
1441 const char *path, int len,
1442 const struct path_simplify *simplify)
1443{
49dc2cc2
RS
1444 struct strbuf sb = STRBUF_INIT;
1445 int baselen, rc = 0;
48ffef96
JH
1446 const char *cp;
1447
1448 while (len && path[len - 1] == '/')
1449 len--;
1450 if (!len)
1451 return 1;
1452 baselen = 0;
1453 while (1) {
1454 cp = path + baselen + !!baselen;
1455 cp = memchr(cp, '/', path + len - cp);
1456 if (!cp)
1457 baselen = len;
1458 else
1459 baselen = cp - path;
49dc2cc2
RS
1460 strbuf_setlen(&sb, 0);
1461 strbuf_add(&sb, path, baselen);
1462 if (!is_directory(sb.buf))
1463 break;
1464 if (simplify_away(sb.buf, sb.len, simplify))
1465 break;
1466 if (treat_one_path(dir, &sb, simplify,
48ffef96 1467 DT_DIR, NULL) == path_ignored)
49dc2cc2
RS
1468 break; /* do not recurse into it */
1469 if (len <= baselen) {
1470 rc = 1;
1471 break; /* finished checking */
1472 }
48ffef96 1473 }
49dc2cc2
RS
1474 strbuf_release(&sb);
1475 return rc;
48ffef96
JH
1476}
1477
dba2e203 1478int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
9fc42d60 1479{
725b0605 1480 struct path_simplify *simplify;
b4189aa8 1481
dba2e203 1482 if (has_symlink_leading_path(path, len))
725b0605
JH
1483 return dir->nr;
1484
1485 simplify = create_simplify(pathspec);
48ffef96
JH
1486 if (!len || treat_leading_path(dir, path, len, simplify))
1487 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1488 free_simplify(simplify);
453ec4bd 1489 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1490 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1491 return dir->nr;
1492}
c91f0d92 1493
686a4a06 1494int file_exists(const char *f)
c91f0d92 1495{
686a4a06 1496 struct stat sb;
a50f9fc5 1497 return lstat(f, &sb) == 0;
c91f0d92 1498}
e6636747
JS
1499
1500/*
9b125da4
NTND
1501 * Given two normalized paths (a trailing slash is ok), if subdir is
1502 * outside dir, return -1. Otherwise return the offset in subdir that
1503 * can be used as relative path to dir.
e6636747 1504 */
9b125da4 1505int dir_inside_of(const char *subdir, const char *dir)
e6636747 1506{
9b125da4 1507 int offset = 0;
e6636747 1508
9b125da4 1509 assert(dir && subdir && *dir && *subdir);
e6636747 1510
9b125da4 1511 while (*dir && *subdir && *dir == *subdir) {
e6636747 1512 dir++;
9b125da4
NTND
1513 subdir++;
1514 offset++;
490544b1 1515 }
9b125da4
NTND
1516
1517 /* hel[p]/me vs hel[l]/yeah */
1518 if (*dir && *subdir)
1519 return -1;
1520
1521 if (!*subdir)
1522 return !*dir ? offset : -1; /* same dir */
1523
1524 /* foo/[b]ar vs foo/[] */
1525 if (is_dir_sep(dir[-1]))
1526 return is_dir_sep(subdir[-1]) ? offset : -1;
1527
1528 /* foo[/]bar vs foo[] */
1529 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1530}
1531
1532int is_inside_dir(const char *dir)
1533{
b892913d
NTND
1534 char cwd[PATH_MAX];
1535 if (!dir)
1536 return 0;
1537 if (!getcwd(cwd, sizeof(cwd)))
1538 die_errno("can't find the current directory");
1539 return dir_inside_of(cwd, dir) >= 0;
e6636747 1540}
7155b727 1541
55892d23
AP
1542int is_empty_dir(const char *path)
1543{
1544 DIR *dir = opendir(path);
1545 struct dirent *e;
1546 int ret = 1;
1547
1548 if (!dir)
1549 return 0;
1550
1551 while ((e = readdir(dir)) != NULL)
1552 if (!is_dot_or_dotdot(e->d_name)) {
1553 ret = 0;
1554 break;
1555 }
1556
1557 closedir(dir);
1558 return ret;
1559}
1560
ae2f203e 1561static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1562{
a0f4afbe 1563 DIR *dir;
7155b727 1564 struct dirent *e;
ae2f203e 1565 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1566 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1567 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1568 unsigned char submodule_head[20];
7155b727 1569
a0f4afbe 1570 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1571 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1572 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1573 if (kept_up)
1574 *kept_up = 1;
a0f4afbe 1575 return 0;
ae2f203e 1576 }
a0f4afbe 1577
ae2f203e 1578 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1579 dir = opendir(path->buf);
c844a803 1580 if (!dir) {
ae2f203e 1581 /* an empty dir could be removed even if it is unreadble */
c844a803
JH
1582 if (!keep_toplevel)
1583 return rmdir(path->buf);
1584 else
1585 return -1;
1586 }
7155b727
JS
1587 if (path->buf[original_len - 1] != '/')
1588 strbuf_addch(path, '/');
1589
1590 len = path->len;
1591 while ((e = readdir(dir)) != NULL) {
1592 struct stat st;
8ca12c0d
AP
1593 if (is_dot_or_dotdot(e->d_name))
1594 continue;
7155b727
JS
1595
1596 strbuf_setlen(path, len);
1597 strbuf_addstr(path, e->d_name);
1598 if (lstat(path->buf, &st))
1599 ; /* fall thru */
1600 else if (S_ISDIR(st.st_mode)) {
ae2f203e 1601 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727
JS
1602 continue; /* happy */
1603 } else if (!only_empty && !unlink(path->buf))
1604 continue; /* happy, too */
1605
1606 /* path too long, stat fails, or non-directory still exists */
1607 ret = -1;
1608 break;
1609 }
1610 closedir(dir);
1611
1612 strbuf_setlen(path, original_len);
ae2f203e 1613 if (!ret && !keep_toplevel && !kept_down)
7155b727 1614 ret = rmdir(path->buf);
ae2f203e
JH
1615 else if (kept_up)
1616 /*
1617 * report the uplevel that it is not an error that we
1618 * did not rmdir() our directory.
1619 */
1620 *kept_up = !ret;
7155b727
JS
1621 return ret;
1622}
039bc64e 1623
ae2f203e
JH
1624int remove_dir_recursively(struct strbuf *path, int flag)
1625{
1626 return remove_dir_recurse(path, flag, NULL);
1627}
1628
039bc64e
JH
1629void setup_standard_excludes(struct dir_struct *dir)
1630{
1631 const char *path;
dc79687e 1632 char *xdg_path;
039bc64e
JH
1633
1634 dir->exclude_per_dir = ".gitignore";
1635 path = git_path("info/exclude");
dc79687e
HKNN
1636 if (!excludes_file) {
1637 home_config_paths(NULL, &xdg_path, "ignore");
1638 excludes_file = xdg_path;
1639 }
69660731 1640 if (!access_or_warn(path, R_OK))
039bc64e 1641 add_excludes_from_file(dir, path);
69660731 1642 if (excludes_file && !access_or_warn(excludes_file, R_OK))
039bc64e
JH
1643 add_excludes_from_file(dir, excludes_file);
1644}
4a92d1bf
AR
1645
1646int remove_path(const char *name)
1647{
1648 char *slash;
1649
1650 if (unlink(name) && errno != ENOENT)
1651 return -1;
1652
1653 slash = strrchr(name, '/');
1654 if (slash) {
1655 char *dirs = xstrdup(name);
1656 slash = dirs + (slash - name);
1657 do {
1658 *slash = '\0';
3fc0d131 1659 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1660 free(dirs);
1661 }
1662 return 0;
1663}
1664
86e4ca69
NTND
1665static int pathspec_item_cmp(const void *a_, const void *b_)
1666{
1667 struct pathspec_item *a, *b;
1668
1669 a = (struct pathspec_item *)a_;
1670 b = (struct pathspec_item *)b_;
1671 return strcmp(a->match, b->match);
1672}
1673
0602f3e9
NTND
1674int init_pathspec(struct pathspec *pathspec, const char **paths)
1675{
1676 const char **p = paths;
1677 int i;
1678
1679 memset(pathspec, 0, sizeof(*pathspec));
1680 if (!p)
1681 return 0;
1682 while (*p)
1683 p++;
1684 pathspec->raw = paths;
1685 pathspec->nr = p - paths;
1686 if (!pathspec->nr)
1687 return 0;
1688
1689 pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1690 for (i = 0; i < pathspec->nr; i++) {
1691 struct pathspec_item *item = pathspec->items+i;
1692 const char *path = paths[i];
1693
1694 item->match = path;
1695 item->len = strlen(path);
8c6abbcd 1696 item->flags = 0;
971e829c
JH
1697 if (limit_pathspec_to_literal()) {
1698 item->nowildcard_len = item->len;
1699 } else {
1700 item->nowildcard_len = simple_length(path);
1701 if (item->nowildcard_len < item->len) {
1702 pathspec->has_wildcard = 1;
1703 if (path[item->nowildcard_len] == '*' &&
1704 no_wildcard(path + item->nowildcard_len + 1))
1705 item->flags |= PATHSPEC_ONESTAR;
1706 }
8c6abbcd 1707 }
0602f3e9 1708 }
86e4ca69
NTND
1709
1710 qsort(pathspec->items, pathspec->nr,
1711 sizeof(struct pathspec_item), pathspec_item_cmp);
1712
0602f3e9
NTND
1713 return 0;
1714}
1715
1716void free_pathspec(struct pathspec *pathspec)
1717{
1718 free(pathspec->items);
1719 pathspec->items = NULL;
1720}
823ab40f
JK
1721
1722int limit_pathspec_to_literal(void)
1723{
1724 static int flag = -1;
1725 if (flag < 0)
1726 flag = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
1727 return flag;
1728}
a39b15b4 1729
270be816
AS
1730/*
1731 * Frees memory within dir which was allocated for exclude lists and
1732 * the exclude_stack. Does not free dir itself.
1733 */
1734void clear_directory(struct dir_struct *dir)
1735{
1736 int i, j;
1737 struct exclude_list_group *group;
1738 struct exclude_list *el;
1739 struct exclude_stack *stk;
1740
1741 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1742 group = &dir->exclude_list_group[i];
1743 for (j = 0; j < group->nr; j++) {
1744 el = &group->el[j];
1745 if (i == EXC_DIRS)
1746 free((char *)el->src);
1747 clear_exclude_list(el);
1748 }
1749 free(group->el);
1750 }
1751
1752 stk = dir->exclude_stack;
1753 while (stk) {
1754 struct exclude_stack *prev = stk->prev;
1755 free(stk);
1756 stk = prev;
1757 }
1758}